This project has three main goals. 1. Analysis of startle modulation. In zebrafish, as in mammals, startle responses are inhibited when the startle stimulus is preceded by a weak auditory prepulse. This form of startle modulation, termed prepulse inhibition, is diminished in neurological conditions including schizophrenia. We conducted a circuit breaking screen using a brain selective enhancer trap library of Gal4 lines to ablate defined groups of neurons before testing PPI. Using this approach we determined that the transcription factor Gsx1 defines neurons required for PPI in both zebrafish and mice. Using volumetric calcium imaging and intersectional genetic approaches, we then located the precise cluster of neurons that regulate prepulse inhibition. Selective optogenetic activatation of these neurons suppresses startle responses to an acoustic stimulus, confirming that this group of neurons control prepulse inhibition. Direct imaging of neurotransmitter release revealed reduced sensory drive from auditory afferents during prepulse inhibition, indicating a central role for presynaptic inhibition in this phenomenon. As Gsx1 neurons only control prepulse inhibition at intervals between prepulse and startle stimulus of greater than 100 ms, we performed a new circuit breaking screen and identified a transgenic line that labels neurons required for prepulse inhibition at short interstimulus intervals. 2. Functional mapping of neuronal architecture mediating short-term behavioral states. We previously demonstrated that light sensitive neurons in the preoptic area of the hypothalamus induce a state of hyperactivity in response to loss of illumination. Behavioral tests demonstrated that this state is part of a light-search strategy. Upon loss of light, larvae initially perform an area-restricted search for illumination. If no light is detected, larvae then swim in an outwards pattern to locate remote sources of light. By performing CRISPR-mediated gene inactivation, we demonstrated that the transition between these two light-search behavioral states requires somatostatin signaling. 3. Development of new tools for analysis of neural circuits. We previously developed a high throughput method for rapidly imaging the entire larval zebrafish brain, then registering images to a reference. Using this method, we developed a brain atlas of transgenic lines, with a focus on Gal4 and Cre lines that can be used for neural circuit analysis. Next, we implemented a computational approach for automated neuroanatomical analysis of the brain that allows us to systematically screen genetic mutant fish for changes in brain microstructure and composition, in an unbiased fashion. We have generated CRISPR mutants for the zebrafish homologs of high impact schizophrenia and autism genes, and compared brain structure in mutants to wildtype fish, revealing previously unsuspected alterations.